json-sKema is a Json Schema validator library for the Java Virtual Machine. It implements the draft 2020-12 specification.
Are you new to JSON Schema? Get started with Understanding JSON Schema!
Add the following dependency to the <dependencies>
section of your project:
<dependency>
<groupId>com.github.erosb</groupId>
<artifactId>json-sKema</artifactId>
<version>0.18.0</version>
</dependency>
dependencies {
implementation("com.github.erosb:json-sKema:0.18.0")
}
// parse the schema JSON as string
JsonValue schemaJson = new JsonParser("""
{
"$schema": "https://json-schema.org/draft/2020-12/schema",
"type": "object",
"properties": {
"age": {
"type": "number",
"minimum": 0
},
"name": {
"type": "string"
},
"email": {
"type": "string",
"format": "email"
}
}
}
""").parse();
// map the raw json to a reusable Schema instance
Schema schema = new SchemaLoader(schemaJson).load();
// create a validator instance for each validation (one-time use object)
Validator validator = Validator.create(schema, new ValidatorConfig(FormatValidationPolicy.ALWAYS));
// parse the input instance to validate against the schema
JsonValue instance = new JsonParser("""
{
"age": -5,
"name": null,
"email": "invalid"
}
""").parse();
// run the validation
ValidationFailure failure = validator.validate(instance);
// print the validation failures (if any)
System.out.println(failure);
// HTTP(s) protocol is also supported
Schema schema = SchemaLoader.forURL("classpath:///path/to/your/schema.json").load();
// create a validator instance for each validation (one-time use object)
Validator validator = Validator.forSchema(schema);
// ...
// Creating a SchemaLoader config with a pre-registered schema by URI
SchemaLoaderConfig config = createDefaultConfig(Map.of(
// When the loader sees this URI,
new URI("urn:uuid:d652a438-9897-4160-959c-bbdb690c3e0d"),
// then it will resolve it to this schema json
"""
{
"$defs": {
"ItemType": {
"type": "integer",
"minimum": 0
}
}
}
"""
));
// parsing the schema json, with a $ref to the above pre-configured URI
JsonValue schemaJson = new JsonParser("""
{
"type": "array",
"items": {
"$ref": "urn:uuid:d652a438-9897-4160-959c-bbdb690c3e0d#/$defs/ItemType"
}
}
""").parse();
// loading the schema json into a Schema object
Schema schema = new SchemaLoader(schemaJson, config).load();
// running the validation
ValidationFailure result = Validator.forSchema(schema).validate(new JsonParser("[null]").parse());
System.out.println(result.toJSON());
If you want to take advantage of the "readOnly"
and "writeOnly"
keywords of json schema, you can tell the Validator
instance if the validation happens in read or write context:
JsonValue schemaJson = new JsonParser("""
{
"type": "object",
"properties": {
"id": {
"readOnly": true,
"type": "number"
},
"name": {
"type": "string"
},
"password": {
"type": "string",
"writeOnly": true
}
}
}
""").parse();
// map the raw json to a reusable Schema instance
Schema schema = new SchemaLoader(schemaJson).load();
// creating write-context validator, it will report validation failures
// for read-only properties that are included in the instance
var writeContextValidator = Validator.create(schema, ValidatorConfig.builder()
.readWriteContext(ReadWriteContext.WRITE)
.build()
);
// creating the json document which will be validated (first in write context, then in read context)
JsonValue instance = new JsonParser("""
{
"id": 1,
"name": "John Doe",
"password": "v3ry_s3cur3"
}
""").parse();
var writeContextFailure = writeContextValidator.validate(instance);
// prints failure because the read-only property "id" is present in write context
System.out.println(writeContextFailure);
// creating read-context validator, it will report validation failures
// for write-only properties that are included in the instance
var readContextValidator = Validator.create(schema, ValidatorConfig.builder()
.readWriteContext(ReadWriteContext.READ)
.build()
);
var readContextFailure = readContextValidator.validate(instance);
// prints failure because the write-only property "password" is present in read context
System.out.println(readContextFailure);
The library allows dynamic schema construction via its SchemaBuilder
class. This is useful in cases when, instead of working with
static schemas or externally fed JSON Schemas, the application needs to create a schema instance at run-time, dynamically. This can
be done with a fluent API, demonstrated below:
import static com.github.erosb.jsonsKema.SchemaBuilder.*;
//...
typeObject()
.property("myString", typeString()
.minLength(2)
.maxLength(50)
.pattern("^[a-zA-Z ]*$")
).property("myInteger", typeInteger()
.minimum(10)
.maximum(20)
.multipleOf(4)
).property("myObject", typeObject()
.additionalProperties(falseSchema())
.patternProperties(Map.of(
"^[A-Z]{2}$", allOf(
typeArray().items(typeNumber())
.prefixItems(List.of(typeBoolean(), typeBoolean())),
typeArray().minContains(3, constSchema(new JsonNumber(1.5)))
)
))
)
The library implements the JSON Schema draft 2020-12 core and validation specifications, with the following notes:
$dynamicAnchor
and$dynamicRef
support is partially implemented
The library currently has built-in support for the following "format"
values defined in the specification:
"format" | Supported? | Notes |
date | Yes | |
date-time | Yes | Non-UTC values with leap seconds not supported |
time | Yes | |
duration | Partially | |
Yes | IPV6 domain parts not supported | |
uri | Yes | |
ipv4 | Yes | |
ipv6 | Yes | |
uuid | Yes |
The following formats are NOT supported: hostname, idn-email, idn-hostname, iri, iri-reference, json-pointer, regex, relative-json-pointer, uri-reference, uri-template .
This project is the successor of everit-org/json-schema. If you want to use draft-04, draft-06 or draft-07 versions of JSON Schema, then you can use the everit library.
Local environment setup:
Prerequisite: JDK and Maven installed
git clone https://github.com/erosb/json-sKema.git
cd json-sKema
git submodule init
git submodule update
mvn clean package
Test annotated with @Tag("acceptance")
require the test suite to be pulled using:
git submodule update --init --recursive
Then run the tests:
mvn clean verify
mvn clean package -Dgroups='!acceptance'